State of the art of smoke and fire detection using image processing

Abstract

In this paper we present a comprehensive review of the state of the art of smoke and fire detection techniques using image processing. Smoke is a good indicator of a pre-fire condition and many fires are indicators of subsequent dangerous situations due to the spread of fire. In this paper we first start our comparison of smoke detection methods and different types of approaches for the classification of smoke. Furthermore we analyse different types of technologies and various models involve in detection techniques such as RGB and HSI models for detecting smoke and fire. Generally, the false alarm rate can be reduced by image processing through effective types of detection techniques such as vision-based or sensor-based methods. Mainly in this paper, we focus on optimised technologies in order to detect smoke and fire at the earliest possible stage of the event, and smoke and fire detection by satellite vision methods.